2D-Leap-Frog and Removal of Outliers in Noisy Photometric Stereo with Non-distant Illuminations

摘要

This paper discusses the reconstruction of a Lambertian surface S_L in three-image noisy photometric stereo under the assumption that light-sources are not necessarily positioned at infinity. The corresponding multi-variable non-linear optimization task either incorporating or not an image boundary continuity enforcement (to remove outliers) is introduced. In addition, a feasible numerical scheme called 2D Leap-Frog is used to recover S_L from three noisy images. The entire setting is tested for non-distant and distant illuminations. The comparison tests are conducted for different surfaces.